Static Cyclic Response of Partially Grouted Masonry. Shear Walls

Transcription

1 Static Cyclic Response of Partially Grouted Masonry Shear Walls Shawn M. Nolph 1 ; Mohamed A. ElGawady 2, M.ASCE Abstract: This research investigated the shear behavior of five full scale partially grouted masonry shear walls (PG- MWs). The walls were built using concrete masonry units and having shear reinforcement ratios ranging from.85% to.169%. The specimens had grout horizontal spacings ranging from 61 mm (24 in.) to 1219 mm (48 in.). All the specimens were tested under constant gravity load and incrementally increasing in-plane loading cycles. In addition, the current provisions of the Masonry Standards Joint Committee (MSJC), New Zealand Code for Masonry Structures, Fattal s model, and strut and tie models were used to predict the shear strengths of the tests specimens. This research showed that there appears to be a maximum shear reinforcement ratio after which no additional shear capacity is achieved. Based on the experimental results, the maximum value appears to be approximately.1% for specimens with a 1219 mm (48 in.) grout horizontal spacing. Increasing the shear reinforcement beyond this level did not increase the shear strength of the test specimens. Finally, the current MSJC shear equations over-estimated the strength of PG-MWs with 1219 mm (48 in.) grout horizontal spacing. A significant source of this error is from over-estimating the contribution of the shear reinforcement. In addition, the current MSJC equations overestimated the masonry contribution to the shear strength of PG-MWs. For partially grouted walls with grout horizontal spacing 813 mm (32 in.), or less, and a horizontal reinforcement ratio of.85%, the MSJC shear equations are adequate. Shear equations by other codes and researchers were unconservative, as well. The strut and tie models were able to predict the shear strength of the test specimens within ±1%. 1 Graduate student, Civil and Environmental Engineering Department, Washington State University, Pullman, WA 99163, s_nolph@wsu.edu 2 Assistant Professor, Civil and Environmental Engineering Department, Washington State University, Pullman, WA 99163, melgawady@wsu.edu 1

2 Key Words: Concrete Masonry; reinforced masonry; cyclic loads; shear walls; shear strength; partially grouted; seismic response Shear Strength of Partially Grouted Masonry Walls Partially grouted masonry wall (PG-MW) is a common form of construction throughout the Unites States, including seismic zones. Most of the currently available test results on PG-MWs are credited to Matsumura (1986) who tested 51 PG-MWs and developed empirical expressions to calculate the shear strength of both fully grouted and partially grouted masonry shear walls. These equations assumed that the contribution of horizontal shear reinforcement and masonry to shear strength of PG-MWs is approximately 6% of those of fully grouted walls. Very few studies have been carried out in the United States to address the in-plane shear strength of PG- MWs. Yancey and Scribner (1989) tested ten PG-MWs under cyclic loading. Horizontal reinforcement ratio ( h ) ranging from to.215% and different distributions of horizontal shear reinforcement were investigated. In addition, the distribution of the shear reinforcement had insignificant effect on the shear strength and ductility. Fattal (1993a) investigated available literature and experimental results on PG-MW and showed that Matsumura s equations predicted lateral strengths that varied from 23% to 18% of the measured strengths of the test specimens with a coefficient of variation of approximately 33%. Fattal (1993a) improved Matsumura s equations and assumed that the contribution of the horizontal shear reinforcement and masonry to the shear strength of PG-MWs are approximately 64% and 8% of those of fully grouted walls, respectively. Using this new set of equations the predicted shear strengths ranged from 41% to 146% of the measured shear strengths with a coefficient of variation of approximately 22%. In another study, Fattal (1993b) analyzed the test data of PG-MWs available in the literature and found that the lateral strengths of PG-MWs increased with increasing horizontal reinforcement ratio, but there is a threshold horizontal reinforcement ratio of.2%, beyond which additional horizontal reinforcement did not appear to be as effective in increasing the lateral strength. Ghanem et al. (1992 and 1993) investigated six one-third scale partially grouted shear walls under in-plane monotonic loading. The effects of reinforcement distribution and applied axial load on shear strength were investigated. All the specimens had a constant horizontal and vertical reinforcement ratio of.12%. The distribution of horizontal and vertical reinforcement as well as the level of the applied axial force had significant effects on the mode of failure, strength, and displacement ductility. As the vertical and horizontal reinforcement were more uniformly distributed, both the strength and ductility of the walls improved. Schultz (1996) 2

3 and Schultz et al. (1998) tested twelve PG-MWs, having aspect ratios of.5,.7, and 1., to examine the effects of distribution and ratio of horizontal reinforcement. It was shown that increasing the aspect ratio from.5 to.7 significantly increased the ultimate shear stresses. However, increasing the aspect ratio from.7 to 1. had a negligible effect on the ultimate shear stresses. In addition, increasing the shear reinforcement ratio was found to increase the ultimate strength of slender walls but decrease those of the squat walls. Finally, it was concluded that the failure mechanism of PG-MWs differs from that of fully grouted shear walls. PG-MWs behaved similar to masonry infilled reinforced concrete frames. Recently, four partially grouted masonry shear walls have been tested (Minaie et al. 21). It was found that the MSJC (28) shear design equations overestimated the shear strengths of the four test specimens by a factor of approximately two. Maleki (28) tested five PG-MWs built according to the Canadian Standard (CSA 24). Out of them, four walls failed in shear while the fifth wall failed in flexure. All the walls had a shear reinforcement ratio ( h ) of.5% and vertical reinforcement ratio of.18%. The horizontal distance between the vertical grouted cells ranged from.56 m to 1.7 m. The constructed walls had two aspect ratios (height/width), 1. and.5, and were tested under zero normal force. The MSJC (28) shear design provisions provided a good prediction of the shear strengths of the test specimens. Shear strength Provisions of MSJC (28) The MSJC (28) provisions use expressions 1 through 5 for calculating the nominal shear strength of masonry shear walls (V n ). These expressions recognize the contribution of masonry (V m ) and horizontal shear reinforcement (V S ). Unlike Matsumura s and Fattal s expressions, the MSJC shear provisions do not distinguish between PG-MW and fully grouted masonry walls. (1) (2) (3) 3

4 (4) (5) Where M/Vd = effective moment to shear ratio, A n = net cross sectional area, f m = specified compressive strength of masonry, P = applied axial load, A sh = cross sectional area of shear reinforcement, and f yh = yield strength of horizontal shear reinforcement. Research Objectives and Methods The current MSJC shear design equations were developed and verified using research on fully grouted masonry shear walls. Shear design of partially grouted masonry walls (PG-MWs) according to MSJC (28) provisions may lead to unsafe shear walls (Minaie et al. 21). However, limited experimental work has been carried out in the United State to investigate the shear strength of partially grouted shear walls. Moreover, some of this data was generated using small scale tests or construction details different from those used in moderate to high seismic zones. Finally, the available experimental data was not well documented and some details of the test specimens were not carefully reported (ElGawady and Elmaprouk 28). The potential issue with the design of PG-MW according to MSJC (28) is a life safety issue which requires immediate attention. This study aims to (a) quantify the effects of the main parameters, namely, the grout horizontal spacing and horizontal shear reinforcement ratio on the performance of PG-MWs, and (b) Investigate the accuracy of the current MSJC (28) shear design provisions. Experimental Program Test specimens Test wall specimens were constructed on oversized heavily reinforced foundations. Professional masons constructed the test specimens in a running bond using standard hollow concrete masonry units (CMUs) and face shell bedding. All specimens were 14 courses high and 6½ block units in length (Fig. 1). Each specimen was nominally 2845 mm (112 in.) high, 2631 mm (13.6 in.) long, and 23 mm (8 in.) wide. Three different configurations of grout horizontal spacing were used as shown in Fig. 2. Specimen identifiers were assigned using the following pattern: PGSSS-GG where PG stands for partially grouted, SSS represents the horizontal reinforcement 4

5 ratio, and GG is the grout horizontal spacing in inches. For example, specimen PG85-48 is a partially grouted specimen having a horizontal shear reinforcement ratio of.85% and a grout horizontal spacing of 1219 mm (48 in.). Table 1 lists the specimen identifiers with the corresponding horizontal shear reinforcement ratios and grout horizontal spacing. Continuous vertical flexural reinforcement was provided in each test specimens, i.e. there was no lap splice for the flexural reinforcement. All specimens had approximately the same total area of flexural reinforcement of 2323 mm 2 (3.6 in. 2 ), corresponding to a vertical reinforcement ratio of.5% (Table 1 and Fig. 2). The flexural reinforcement was selected such that the flexural capacity of every specimen exceeded its predicted shear capacity. The shear capacity of each specimen was calculated using Eqs. 1 to 5. Horizontal shear reinforcement was provided in bond beam knockout blocks placed at a spacing of 1219 mm (48 in.) in the 6 th and 12 th courses. All specimens except specimens PG and PG had 1 D 16 (#5) as shear reinforcement in every bond beam. Specimens PG and PG had 1 D 19 (#6) and 2 D 16 (#5) as shear reinforcement in every bond beam, respectively. The shear rebar was anchored with MSJC (28) provisions-compliant 18-degree hooks around the outermost vertical reinforcement. All specimens were constructed using hollow CMUs having nominal dimensions of 23 mm x 23 mm x 46 mm (8 in. x 8 in. x 16 in.) for full blocks and 23 mm x 23 mm x 23 mm (8 in. x 8 in. x 8 in.) for half blocks. The CMUs had a measured net area compressive strength of 18.1 MPa (263 psi). The walls were built in three consecutive days. Courses 1 through 6 for each specimen were constructed on the first day. On the second day of construction, courses 1 through 6 were grouted in each wall and courses 7 through 12 were built. To connect the lateral load actuator to the wall (as explained later), 19 mm ( 3 / 4 in.) diameter threaded rods were installed in every cell at the 12 th masonry course prior to grouting. On the third day of construction, courses 7 through 12 in each specimen were grouted and courses 13 and 14 were laid and fully grouted. Each specimen was grouted using fine aggregate grout provided by a local ready-mix supplier. The grout had a measured compressive strength f g` = 29.2 MPa (424 psi) (ASTM C119-7). Masonry prisms were constructed during the construction of the wall-specimens and were tested according to ASTM C The masonry compressive strength f m` was 11.3 MPa (164 psi) for ungrouted prisms and 19.7 MPa (286 psi) for grouted prisms. All the rebar used in the construction was Gr. 6 with average measured yield strength of 439 MPa (63.6 ksi). 5

6 Test setup and loading system Fig. 3 illustrates the test setup. A constant vertical force (P u ) of approximately 49.4 kn (11.1 kips) corresponding to an applied axial stress of.1 MPa (14 psi) was applied to the top of each specimen using two hydraulic jacks. The jacks were attached from its top end to a trolley having a minimal coefficient of friction that was free to move laterally under the I-beam girder. The jacks were connected from their bottom ends to a hollow structural steel tube 12 mm x 23 mm x 6 mm (4 in. x8 in.x 1 / 4 in.) which uniformly distributed the applied axial loads to the top surface of each test specimen. The applied axial loads were kept constant during testing. A 89 kn (2 kips) capacity single-ended hydraulic actuator was used to apply the required displacement at the 12 th CMU course of each test specimen. The actuator was attached to a loading steel frame at one end and to a pair of C-channels that were bolted to each wall specimen using thirteen 19 mm ( 3 / 4 in.) threaded rods that were grouted in place in the 12 th CMU course during construction (Fig. 3). Finally, the foundation of each specimen was post-tensioned to the laboratory strong floor with six threaded rods. Instrumentations Each specimen was typically instrumented with 13 strain gages having a gauge length of 6.3 mm (.25 in.) bonded to the flexural and shear reinforcing steel prior to construction of the walls and 17 string potentiometers mounted on the walls and foundations immediately before testing (Fig. 4). The middle and top shear reinforcement had four strain gages each. The applied lateral load was measured using the load cell on the actuator. Experimental Results Specimen PG85-48 Specimen PG85-48 was constructed with 1219 mm (48 in.) grout horizontal spacing and a single D 16 (#5) rebar in each bond beam (ρ h =.85%). The first cracks were stair-step cracks passing through the mortar bed and head-joints in the bottom masonry panels in the south and north directions when pushing and pulling the specimen, respectively, to a drift angle of.11% (4. kn (9. kips)). While pushing toward the north to a drift angle of.27% (85.9 kn (19.3 kips)), a horizontal flexural crack developed in the mortar joint in the south end grouted cell just beneath the middle bond beam. At a drift angle of.33% (98.3 kn (22.1 kips)), stair-step cracks in the mortar bed and head-joints developed in the upper masonry panels. As testing continued, 45 cracks developed through the 6

7 CMUs in all panels and middle bond beam. At a drift angle of 1.3% (221.8 kn (49.9 kips)), several diagonal cracks opened significantly and the peak strength of the wall was achieved. At this drift angle, several loud "pops" and bulging of the face shells of several units were noted. At a drift angle of 1.5% (166.1 kn (37.4 kips)), an approximate 24% drop in the average lateral strength happened. By the end of the test, all masonry panels had stair-step and/or 45 cracks indicative of the formation of compression struts within the panel (Fig. 5). In addition, spalling of the south end shells of the 3 rd through the 5 th CMU courses was observed (Fig. 5). Table 2 summarizes the main events occurred during testing the specimen. Specimen PG12-48 Specimen PG12-48 was constructed with 1219 mm (48 in.) grout horizontal spacing and a single D 19 (#6) rebar in each bond beam (ρ h =.12%). The specimen behaved similar to specimen PG First, stair-step cracks appeared in the bottom masonry panel followed by stair-step cracks in the top masonry panels. During testing to a drift angle of.27% (82.7 kn (18.6 kips)), a 45 crack developed through the face shell of the end cell of the southern CMU in the middle bond beam. During pulling to a drift angle of.87% (173.8 kn (39.1 kips)), a vertical splitting crack appeared in the end shell of the southern CMU of the middle bond beam. More 45 cracks developed in the mortar joints and masonry units until a drift angle of 1.5% (227.7 kn (51.2 kips)) at which point the wall reached its peak strength. The vertical splitting crack in the south end shell of the bond beam significantly extended through 4 masonry courses while pulling to a drift angle of 1.5%. By the end of the 3 rd cycle at a drift angle of 1.5%, a significant stair-step crack developed in the wall and the splitting crack widened significantly (Fig. 6). At this cycle, the lateral strength of the wall dropped by an average of approximately 19% and the test was ended at this point. At the termination of the test, the east side face-shells and end-shells of masonry units in courses 3 through 6 were found to be detached from the wall but still resting in place. Fig. 6 shows a close-up of the damage after the detached masonry was removed. Table 2 summarizes the main events occurred during testing the specimen. Specimen PG Specimen PG was constructed with 1219 mm (48 in.) grout horizontal spacing and two D 16 (#5) rebar in each bond beam (ρ h =.169%). Similar to specimens PG12-48 and PG85-48, stair-step cracks appeared in the lower masonry panels followed by the upper masonry panels. Then, during testing to a drift angle of.54% (185 kn (41.6 kips)), a vertical crack in the face shell of the last unit at the south end of the 7 th course (i.e. just above the 7

8 middle bond beam) developed (Fig. 7). This crack was similar to those observed in specimens PG85-48 and PG12-48 that led to the detachment of masonry face shells. By a drift angle of.65% (22.9 kn (45.7 kips)), the specimen reached its peak lateral strength and the diagonal cracks widths significantly increased. During testing to a drift angle of.87% (158.9 kn (35.7 kips)), an approximate 22% drop occurred in the lateral strength of the specimen and the test was stopped. Table 2 summarizes the main events occurred during testing the specimen. Specimen PG85-32 Specimen PG85-32 was constructed with 813 mm (32 in.) grout horizontal spacing and a single D 16 (#5) in each bond beam (ρ h =.85%). The first cracks were stair-step crack passing through the outermost bottom masonry panels. These cracks were followed by stair-step crack in the top masonry panels. By reaching a drift angle of.87% (213 kn (47.8 kips)), 45 cracks developed through the CMUs and bed and head-joints in all panels (Fig. 7). The specimen reached its peak lateral load at a drift angle of 1.3% (26. kn (58.5 kips)) when a significant diagonal crack passing through CMUs and mortar joints developed along the full diagonal length of the upper south masonry panel. As the applied lateral drift angle increased to 1.5% (21.4 kn (45.3 kips)), this diagonal crack extended and became a vertical splitting crack on the face shells of CMUs and passing through the middle bond beam all way through the 3 rd brick course (Fig. 7). By this drift angle, the test was stopped where significant 45 cracks passing through the CMUs and mortar joints developed leading to an approximate reduction of the lateral strength of the specimen by 23%. Table 2 summarizes the main events occurred during testing the specimen. Specimen PG85-24 Specimen PG85-24 was constructed with 61 mm (24 in.) grout horizontal spacing and a single D 13 (#5) in each bond beam (ρ h =.85%). Similar to the other specimens, stair-step cracks developed in the lower masonry panels followed by the upper masonry panels. Then, 45 cracks developed through the CMUs in all panels. While testing to a drift angle of 1.1% (24 kn (54. kips)), diagonal cracks formed in the face shells in the bond beam and in the upper southern masonry panel (Fig. 8). While testing to a drift angle of 1.3% (6.5 kips (269 kn)), a vertical splitting crack on the south end formed in the end shell at the 6 th course. While testing to a drift angle of 1.5% (66.4 kips (295 kn)), the specimen reached its peak strength and the diagonal cracks, that started at a drift angle of 1.1%, extended down into the lower southern panel. These cracks opened significantly while testing to a drift angle of 1.7% (54.2 kips (241 kn)) leading to reduction in the lateral strength of the specimen by an average of 19%. In addition, 8

9 the splitting crack in the south end shells extended along the height of four masonry units. Table 2 summarizes the main events occurred during testing the specimen. Hysteretic Performance of The Test Specimens The load-drift hysteretic curves of the test specimens are shown in Fig. 9. The hysteretic response was obtained by plotting the measured lateral forces from the load cell of the actuator versus the measured displacement at 12 th masonry course. As shown in the figure, the hysteretic behaviors of all specimens were similar except for specimen PG where the high reinforcement ratio led to brittle failure. All specimens displayed stable symmetrical hysteresis loops with relatively narrow loops before reaching its ultimate lateral strength at lateral drifts of approximately 1.1 to 1.5%, except for specimen PG Specimen PG169-48, which had the highest shear reinforcement ratio, reached its ultimate lateral strength at a lateral drift angle of approximately.65%. All specimens behaved approximately linear elastic until lateral drifts of.5 to.11% where the first shear crack occurred. Beyond that the stiffness of the specimens degraded with the specimens still able to carry the applied lateral and vertical forces. Once the specimen started the nonlinear inelastic behavior the residual drift angle values increased. At the end of the tests the average residual drift angle value was approximately 35% of the applied peak lateral drift angle. Once the specimens reached their peak strengths, the strengths degraded very quickly and testing was terminated when the lateral resistance of the specimens dropped by approximately 2% of the peak strength. A test specimen was considered to have reached failure at the 2% drop from peak lateral strength. The specimens failed at lateral drifts of approximately 1.3 to 1.7% except for specimen PG which failed at a lateral drift angle of.87%. Discussion This section discusses the behavior of the test specimens. The discussion includes the following characteristics of the test specimens: (1) Effects of grout horizontal spacing; (2) Effects of horizontal shear reinforcement ratio; and (3) Comparisons between the measured and calculated strengths of the test specimens. Effects of Grout Horizontal Spacing The effects of grout horizontal spacing on lateral strength, stiffness, and displacement ductility is presented in this section. Grout horizontal spacings were 1219 mm (48 in.) (specimen PG85-48), 813 mm (32 in.) (specimen 9

10 PG85-32), and 61 mm (24 in.) (specimen PG85-24). These three specimens had the same horizontal reinforcement ratio of.85%. Fig. 1(a) shows the backbone curves of these three specimens obtained using the procedure detailed in FEMA 356 (2). Fig. 1(b) shows the backbone curves of these specimens after normalizing by the peak strength of each specimen. Examination of Fig. 1 shows the effects of the grout horizontal spacing on the lateral strength, stiffness, and displacement ductility of these three test specimens. The grout horizontal spacing did not have systematic effects on the deformability of the test specimens. All test specimens reached approximately the same ultimate drift angles with specimen PG85-24 having slightly higher ultimate drift angle. Similarly, there were no systematic effects of the grout horizontal spacing on the displacement ductility of the test specimens. The three partially grouted specimens reached a displacement ductility factor of approximately The displacement ductility of each specimen was calculated using an idealized elasto-plastic backbone curve (Nolph 21). Finally, the three test specimens had approximately the same initial stiffness with specimen PG85-48 having a slightly lower initial stiffness. Fig. 11 shows the effect of grout horizontal spacing as well as the net cross sectional area on the shear strengths of the test specimens. As shown in the figures, specimen PG85-24 has the highest strength followed by specimens PG85-32, and PG By increasing the grout horizontal spacing or decreasing the net cross sectional area the strength of the test specimens linearly decreased. Fig. 12 shows the lateral drift angle versus the net shear stresses on the same set of specimens. At each drift level, the shear stress was calculated as the average of the peak lateral forces at that drift level divided by the net cross sectional area of each specimen considering face-shell bedding. As shown in the figure, all specimens were able to carry approximately the same peak shear stresses. It is worth noting that the provisions of the current Masonry Standards Joint Committee (MSJC 28) use the net cross sectional area for calculating the shear strength of partially grouted walls as shown in Eq. 2. However, other international building standards such as the current New Zealand standard, NZS 423 (24), uses the cross sectional area corresponding to the thickness of the face shells as the net cross sectional area in the case of partially grouted walls. This criterion was selected to satisfy the shear flow continuity requirements and to avoid the potentials of vertical shear failure of continuous ungrouted cells (Voon 27). Figure 12(b) shows the shear stress calculated according to NZS 423 (24) versus the lateral drift for the same set of specimens. Using the NZS 423 (24) recommendations, the ultimate shear stresses in the partially grouted specimens varied depending on the grouting scheme. This shows that using the net cross sectional area as 1

11 defined by the current MSJC provisions is more appropriate than the net cross sectional area as defined by the NZS 423 (24). As mentioned earlier, the axial strains in the rebar in the middle bond beams were measured using strain gages. The lateral force vs. axial strain in the shear reinforcement is shown in Figure 13. In addition, vertical lines representing the value of the rebar yield strain are shown in the same figure. As shown in the figure the response was stable and symmetric. For small applied lateral forces, the permanent dilation in the rebar (measured at zero lateral force) is small indicating minimal opening of shear cracks. Once the applied lateral force increased, the rate of increase in the axial strains in the rebar increased and the residual strains at zero lateral force increased indicating increase in the shear crack widths. For specimens having the same horizontal reinforcement ratio but different grout horizontal spacing, the rebar in the three specimens reached or exceeded the steel yield strain. Interestingly, using smaller horizontal spacing between the vertical grouted cells resulted in higher strains in the shear rebar. For specimen PG85-48, the shear rebar reached a peak axial strain of 26 micro-strain which is approximately 119% of the yield strain of the rebar. The shear rebar reached its yield strain just few loading cycles before the specimen reached its ultimate displacement and by the end of the test, the residual strain in the shear rebar was approximately 7 micro-strains. For smaller grout spacing i.e. 813 mm (32 in.) and 61 mm (24 in.), the ultimate strains were significantly higher than the yield strain of the rebar with high residual strains. The ultimate strains in specimens PG85-32 and PG85-24 were 74, and 7 micro-strain which are approximately 338%, and 32% of the yield strain of the rebar, respectively. By the end of the test, the residual strains in the shear rebar of specimens PG85-32 and PG85-24 were approximately 47, and 5 micro-strains, respectively. The measured strains in the shear rebar were converted into axial forces by multiplying the measured axial strains by the cross sectional area of shear reinforcement rebar and E-modulus of the rebar. The calculated forces in the rebar were capped by the force corresponding to the yield stresses in the rebar. The peak lateral force in the rebar at each drift angle level was plotted vs. the lateral drift angle in Fig. 14. Also, shown in the figure is the lateral drift vs. the applied shear force i.e. backbone curves. The difference between the force in the shear reinforcement and the backbone curve represents the forces carried by masonry alone. As shown in the figure, the specimens exhibited similar patterns of force development in the masonry with the engagement of the shear reinforcement started at a lateral drift of approximately.4%. Before that, the masonry alone is resisting the applied shear force. Finally, the masonry contribution to the shear strength increased with decreasing grout horizontal spacing. 11

12 Effects of Shear Reinforcement Ratio For specimens having grout horizontal spacing of 1219 mm (48 in.), horizontal reinforcement was provided at ratios of.85% (specimen PG85-48),.12% (specimen PG12-48) and.169% (specimen PG169-48). Fig. 15(a) shows the backbone curves of the three specimens that had the same grout horizontal spacing but different reinforcement ratio. Fig. 15(b) shows the backbone curves of the same three test specimens after normalizing by the peak strength of each specimen. Fig. 16(a) shows the average peak strength vs. provided horizontal reinforcement ratio. Examination of Figs. 15 and 16(a) show the effects of the horizontal shear reinforcement ratio on the lateral strength, initial stiffness, and displacement ductility of the test specimens. As shown in the figures, increasing the shear reinforcement from 1 D 16 (#5) in specimen PG85-48 to 1 D 19 (#6) in specimen PG12-48 slightly increased the lateral strength from kn (5 kips) to kn (51.5 kips). However, such increase in the shear strength with increasing the shear reinforcement ratio was not observed in specimen PG The shear reinforcement ratio of specimen PG was double that of specimen PG85-48 but its shear strength was only approximately 92% of that of specimen PG It is worth noting that Voon (27) observed similar behavior for fully grouted walls. When the shear reinforcement was increased by a factor of 2.5, the shear strength remained constant. Voon (27) concluded that for a given masonry wall there is a certain threshold of shear reinforcement beyond which there is no effect from any additional shear reinforcement. For partially grouted walls, Fattal (1993b) suggested an upper limit of.2% for shear reinforcement. This is close to the shear reinforcement provided to specimen PG Finally, changing the horizontal reinforcement ratio neither significantly changed the initial stiffness nor the deformability of specimens PG85-48 and PG However, doubling the shear reinforcement ratio from specimen PG85-48 to PG significantly reduced the ultimate drift angle of specimen PG Fig. 13 shows the axial strains in the shear reinforcement vs. the applied lateral load. As shown in the figure, the behaviors of the three specimens are identical. However, for specimens having high shear reinforcement ratio, i.e. specimens PG and PG12-48, the rebar did not reach the yield strains with ultimate strains of 11 and 18 micro-strains which are approximately 5% and 82% of the yield strain of the rebar, respectively. By the end of the test, the residual strains in the shear rebar were approximately 2 and 5 micro-strains for specimens PG and PG12-48, respectively. For specimen PG85-48, the shear rebar reached a peak axial strain of 26 microstrain which is approximately 119% of the yield strain of the rebar. Fig. 16(b) shows the maximum axial strains in the horizontal reinforcement vs. the shear reinforcement ratio. As the figure shows, increasing the shear 12

13 reinforcement ratio caused a linear decrease in the ultimate strain in the shear reinforcement. Also, shown in the figure is a line representing the relationship between the horizontal reinforcement ratio and the peak strains in the rebar. This relationship was developed based on regression analysis of these limited test data. Based on the regression analysis, a horizontal reinforcement ratio of approximately.11%, or less, will result in yielding of the horizontal reinforcement. As indicated in Eq. 3, the MSJC (28) shear provisions assume yielding of the shear reinforcement which was not the case for horizontal reinforcement ratios greater than approximately.11%. However, this proposed ratio is based on the results of only 3 specimens and more specimens are required to come up with an upper limit for shear reinforcement ratio. Fig. 17 shows the lateral drift vs. the total force in the shear reinforcement in the mid-wall bond beam (6 th CMU course) as well as lateral drift vs. the applied shear force (backbone curve). For specimens PG85-48 and PG12-48, before a lateral drift of approximately.4%, the masonry alone resisted the applied shear force. Beyond that, the shear rebar started to be engaged. For specimen PG169-45, the shear rebar started to be engaged at a lateral drift of approximately.2%. Interestingly, the steel rebar developed approximately the same force contributions in all three specimens. Measured vs. Predicted Shear Strength Using MSJC (28) Provisions The predicted strength of each test specimen (V n ) using the MSJC (28) shear equations (Eqs. 1-5) are presented in Figs 9, 11, and 16. In addition, V n times the shear strength reduction factor ( ) of.8 is presented in Fig.9. The errors in the predictions are shown in Fig. 18. The error is defined as the (ultimate strength V n )/ultimate strength. As shown in the figure, the MSJC (28) shear equations overestimated the shear strengths for all specimens with a single exception, i.e. specimen PG For grout horizontal spacing of 61 mm (24 in.) and 813 mm (32 in.), the predictions were quite good with an over-prediction of 7% for specimen PG85-32 and an underprediction of 1% for specimen PG For the three specimens with 1219 mm (48 in.) grout horizontal spacing and variable shear reinforcement ratios, the MSJC (28) shear equations over-predicted the shear strengths with an error ranged from 12% to 26%. Lines representing the correlation relationship between the test results and grout spacing and reinforcement ratio are presented in Fig. 18. Figs. 16 and 18(a) shows that the error in predicting the shear strength using MSJC (28) provisions increased with increasing reinforcement ratio. 13

14 For the three specimens with reinforcement ratio of.85% and variable grout horizontal spacing, Fig. 18(b) shows that the error in the strength predictions using the MSJC (28) shear equations is linearly correlated to the grout horizontal spacing. With the shear reinforcement ratio constant, V ns (Eq. 3) is constant for all three specimens. Therefore, the total nominal shear strength (Eq. 1) varies due to V nm, the nominal shear strength contributed by the masonry (Eq. 2). The contribution to V nm from the axial load is constant for all three specimens. The only variable between the three specimens is the net cross sectional area of masonry. This suggests that a reduction factor of some type should be applied to the nominal shear strength contributed by the masonry, V nm, for partially grouted shear walls. It is worth noting that both Matsumura s and Fattal s model used a reduction factor applied to the V nm term in the case of partially grouted walls. Eq. 6 shows a reduction factor developed based on the limited data set presented in this manuscript. (6) Where A g is the gross cross sectional area. Other Codes and Methods The strengths of the test specimens were calculated using the New Zealand Code for Masonry Structures NZS 423 (24), Fattal s model (1993a), and strut and tie models (S&T). The S&T models followed the recommendations given in ACI 318 (28) appendix B. The compressive strengths of the fully grouted and ungrouted prisms were used for the compressive strengths at the nodes and compression members, respectively. More details about the different methods are available in Nolph (21). Fig. 19 shows the experimental results and the predictions using the MSJC (28) strength design provisions, the equations by Fattal (1993a), the NZS 423 (24), and the strut and tie models. As shown in the figure, the equations by Fattal (1993a) were unconservative for three specimens, i.e. PG85-48, PG12-48, and PG169-48, but was less unconservative than the MSJC (28). For the remaining specimens, PG85-32 was predicted fairly well (-9% error) but PG85-24 was predicted at only 8% of the experimental strength. The NZS 423 (24) was unconservative for three specimens, i.e. PG85-48, PG12-48, and PG169-48, and was reasonably close to the MSJC (28) predictions for these specimens. Predictions for specimens PG85-32 and PG85-24 were conservative. The NZS 423 (24) predictions were generally similar to those of the MSJC (28) and did not 14

15 present a substantial improvement. The strut and tie model was a good predictor for all specimens with the strengths of the specimens predicted within ±1%. SUMMARY AND CONCLUSIONS This research investigated the shear behavior of five full scale partially grouted masonry shear walls (PG- MWs). The walls were built using concrete masonry units and having shear reinforcement ratios ranged from.85% to.169%. The specimens had grout horizontal spacings ranged from 61 mm (24 in.) to 1219 mm (48 in.). All the specimens were tested under constant gravity load and incrementally increasing in-plane loading cycles. In addition, the current provisions of the Masonry Standards Joint Committee (MSJC), New Zealand Code for Masonry Structures, Fattal s model, and strut and tie models were used to predict the shear strengths of the tests specimens. The following findings and conclusions were drawn from the research presented in this manuscript: There appears to be a maximum shear reinforcement ratio after which no additional shear capacity is achieved. Based on the experimental results, the maximum value appears to be approximately.1% for specimens with a 1219 mm (48 in.) grout horizontal spacing. Increasing the shear reinforcement beyond this level did not increase the shear strength of the test specimens. Beyond a horizontal reinforcement ratio of.85%, i.e. at.12% and.169%, failure in the wall specimens occurred without the shear reinforcement reaching its yield strain. A similar statement cannot be made for the 813 mm (32 in.) and 61 mm (24 in.) grout horizontal spacings due to there being only one shear reinforcement level,.85%, tested at these grout horizontal spacings. Increasing the reinforcement ratio led to a very stiff behavior and limited deformability of the test specimens. Within the scope of the investigated parameter in this research, there seems to be insignificant effects of the grout horizontal spacing on the deformability and the stiffness of the test specimens. The current MSJC shear equations over-estimated the strength of PG-MWs with 1219 mm (48 in.) grout horizontal spacing. A significant source of this error is from over-estimating the contribution of the shear reinforcement. In addition, the MSJC (28) equations overestimated the masonry contribution. For partially grouted walls with grout horizontal spacing 813 mm (32 in.), or less, and a horizontal reinforcement ratio of.85%, the equations are adequate. 15

16 In general, the equations by Fattal (1993a and b) were unconservative but better than those of the current MSJC (28). The New Zealand masonry code predictions were judged to be generally similar to those of the MSJC (28) and did not present a substantial improvement. The strut and tie model was a good predictor for all specimens. The partially grouted specimens were all predicted within ±1%. ACKNOWLEDGMENTS This research was conducted with funding from the National Concrete Masonry Association, the Northwest Concrete Masonry Association, and the Eastern Washington Masonry Promotion Group. Appreciation is also extended to Mr. R. Duncan, and S. Lewis for technical support during construction and testing the specimens. REFERENCES 1. American Concrete Institute: 25, Building Code Requirements for Structural Concrete (ACI 318-5) and Commentary (ACI 318R-5), Farmington Hills, MI. 2. Elmapruk, J.H., ElGawady, M.A., "Evaluation of the MSJC 28 Shear Strength Equations For Partially Grouted Masonry Shear Walls." 11th Canadian Masonry Symposium, Toronto, Ontario, May 31-June 3, Fattal, S. G. (1993a) Strength of partially-grouted masonry shear walls under lateral loads, NISTIR , Gaithersburg, MD. 4. Fattal, S. G. (1993b) The effect of critical parameters on the behavior of partially grouted masonry shear walls under lateral loads, NISTIR , Gaithersburg, MD 5. FEMA 356 (2) Prestandard and Commentary for the Seismic Rehabilitation of Buildings, Federal Emergency Management Agency, Washington, D.C. 6. Ghanem, G. M., (1992), Effect of Steel Distribution on the Behavior of Partially Reinforced Masonry Shear Walls 6 th Canadian Masonry Symposium, University of Saskatchewan, pp Ghanem, G. M., Salama, A. E., Elmagd, S. A., and Hamid A. A. (1993), Effect of Axial Compression on the Behavior of Partially Reinforced Masonry Shear Walls, 6 th NAMC, Philadelphia, pp Maleki, M., (28), Behavior of partially Grouted Reinforced Masonry Shear walls under Cyclic Reversed Loads Thesis of Doctoral of Philosophy, McMaster University. 16

17 9. Masonry Standards Joint Committee: 28, Building Code Requirements for Masonry Structures, TMS 42-8, The Masonry Society, Boulder, CO, ACI 53-8, American Concrete Institute, Farmington Hills, MI, ASCE 5-8, American Society of Civil Engineers, Reston, VA. 1. Matsumura, A. (1987), Shear Strength of Reinforced Hollow Unit Masonry Walls in Proceeding, 4 th North American Masonry Conference, Los Angeles, California pp Minaie, E., Mota, M., Moon, F., and Hamid, A. (21), In-plane behavior of partially grouted reinforced concrete masonry shear walls J. of Structural Engineering, Vol. 136(9), pp New Zealand Standard 423:24, Design of Reinforced Concrete Masonry Structures, Standards Association of New Zealand, Wellington. 13. Nolph, S. (21), In-plane shear performance of partially grouted masonry shear walls MSc. Thesis, Washington State University, Pullman, WA. 14. Schultz, A.E, Hutchinson, R. S., (1998), Seismic Performance of Masonry Walls with Bed Joint Reinforcement Elsevier Science Journal, T Schultz, A.E, (1994), Seismic Resistance of Partially-Grouted Masonry Shear Walls Structural Concrete and Masonry (ASCE Structures Congress XIV) Chicago- IL, pp Voon, K. C., (27), "In-Plane Seismic Design of Concrete Masonry Structures PhD thesis, The University of Auckland, NZ. 17. Yancey, C.W.C., and Scribner, (1989), Influence of Horizontal Reinforcement on Shear Resistance of Concrete Block Masonry Walls NISTER , Maryland 17

18 Figure 1: Typical dimensions of a test specimen [in. (mm)] (a) PG85-48, PG12-48, PG (b) PG85-32 (c) PG85-24 Figure 2: Horizontal cross-sections of the different wall specimens 18

19 Figure 3: Test Setup Figure 4: Typical locations of strain gages and string pots 19

20 North South Figure 5: Specimen PG85-48 at the test end North South Figure 6: Specimen PG12-48 at the test end 2

21 South North North (a) (b) Figure 7: Specimens (a) PG and (b) PG85-32 at the test end South South Figure 8: Specimen PG85-24 at the test end North 21

22 (a) (b) (c) (d) (e) Figure 9: Load-drift hysteretics for specimens (a) PG85-48, (b) PG12-48, (c) PG169-48, (d) PG85-32, and (e) PG

23 (a) (b) Figure 1: Specimens with the same horizontal reinforcement ratio and differing grout horizontal spacing (a) backbone, and (b) normalized backbone curves Figure 11: Effects of grout horizontal spacing on shear strength of the test specimens (a) (b) Figure 12: Drift vs. net shear stress for ρ h =.85%, (a) using net cross sectional area, and (b) using face-shell only cross sectional area 23

24 Load (kip) Load (kn) Load (kip) Load (kn) Load (kip) Load (kn) Load (kip) Load (kn) load (kip) yield microstrain microstrain (a) (b) yield microstrain microstrain (c) (d) yield microstrain (e) Figure 13: Lateral load vs. axial strain in the shear steel for specimens (a) PG85-48, (b) PG12-48, (c) PG169-48, (d) PG85-32, and (e) PG

25 Force (kip) Force (kn) Force (kip) Force (kn) Force (kip) Force (kn) Backbone Shear Steel Backbone Shear Steel Drift (%) Drift (%) (a) (b) Backbone Shear Steel Drift (%) (c) Figure 14: Drift vs. force in shear reinforcement and applied force (backbone curve) for specimens (a) PG85-48, (b) PG85-32, and (c) PG (a) (b) Figure 15: Backbone curves for specimens with 1219 mm (48 in.) grout horizontal spacing and varying horizontal reinforcement ratios. 25

26 Force (kip) Force (kn) Force (kip) Force (kn) Force (kip) Force (kn) (a) (b) Figure 16: Effects of ρ h on (a) shear strength, and (b) maximum strain in shear reinforcement Backbone Shear Steel Backbone Shear Steel Drift (%) Drift (%) (a) (b) Backbone Shear Steel Drift (%) (c) 26

27 Figure 17: Drift vs. force in shear reinforcement and applied force (backbone curve) for specimens (a) PG85-48, (b) PG12-48, and (c) PG (a) (b) Figure 18: Errors in the predicted shear strengths for specimens having (a) different reinforcement ratio, and (b) grout horizontal spacing 27

28 Figure 19: Experimental shear strength and shear strength predictions using MSJC (28), Fattal equations, New Zealand code (NZ), and strut and tie model (S&T). 28

29 Table 1: Wall specimen parameters Wall ID PG85-48 PG12-48 PG PG85-32 PG85-24 Vertical reinforcement spacing in inch (mm) Vertical reinforcement A n in inch 2 (mm 2 ) 48 (1219) 48 (1219) 48 (1219) 32 (813) 24 (61) 2 #7 x 3 cells 2 #7 x 3 cells 2 #7 x 3 cells 2 #6 x 4 cells 2 #6 x 2 cells (229,677) (229,677) (229,677) (25,322) (271,612) shear stirrup 1 #5 1 #6 2 #5 1 #5 1 #5 ρ h A v in inch (mm 2 ) (2) (284) (4) (2) (2) Constant parameters for all specimens: hw = 92 in (2337 mm); t = in (194 mm); dv = 13.6 in (2631 mm); Pu = 118 lb (49286 N); s = 48 in (1219 mm); fy = 636 psi (438.5 MPa); f`m = 164 psi (11.3 MPa), except FG85- where f`m = 286 psi (19.7 MPa). 29

30 Table 2: Test summary Crack description PG85-48 PG12-48 PG PG85-32 PG85-24 First stair-step crack.11% (4. kn).11% (38.7 kn).5% (33.4 kn).11% (39.1 kn).5%* (19.1 kn) First stair-step crack in the.33%.27%.33%.11%.22% upper panel (98.3 kn) (82.7 kn) (98.3 kn) (39.1 kn) (72.1 kn) Ultimate strength 1.3% (27. kn) 1.5% (199 kn).65% (22.9 kn) 1.3% (26 kn) 1.5% (295 kn) End of the test 1.5% (19. kn) 1.5% (161. kn).87% (158.9 kn) 1.5% (237 kn) 1.7% (241 kn) Hair cracked occurred in the bed and head-joints of this specimen developed while moving the specimen into the testing frame. 3